Lying scientists and the lying lies they tell

When you read academic papers, you aren't looking for treachery and deceit behind the stolid prose. Don't be so trusting: universities can be a wretched hive of scum and villainy. Here's help.

canyonoffools07-19-2016-2427.jpg

Don't go there. Really.

Photo by Robin Harris.

Last week in When Big Data is Bad Data, I wrote about how faulty software invalidated some 40,000 fMRI-based papers. But bad software isn't the only problem facing those of us who want evidence-based decisions.

There's immense pressure on academics to publish -- or perish. Inevitably, some take shortcuts to achieve their goals.

Garbage in, garbage out

The online publication Retraction Watch covers the ivy-covered mean streets of academe. Once you get past the PhDs and the polysyllabic jargon, you have to wonder how smart some of these shysters are.

The dude who doctored data on 11 papers?

The Harvard researcher who got his PhD revoked for fudging data? His former group earned three retractions as well.

Or the guy who sued to stop retractions and just got hit with seven more? The "best defense is a good offense" strategy is not working.

Or the guy who faked 70 experiments!

Then there's Yoshitaka Fujii, who reportedly holds the current record for most retractions at 183. Wouldn't it have been easier to just do the work on fewer papers?

Good news

The news isn't all bad.

A new idea, post-publication peer review, is starting to take hold. In-depth stat reviews are time consuming, so focusing on the 10 to 20 percent of papers that get published makes sense.

But as Retraction Watch's own 10 most highly-cited retracted papers list shows, even retracted papers continue to get cited. The paper that falsely linked vaccines with autism has had over 300 citations since retraction.

The Storage Bits take

We can be skeptical of scientific research without being cynical about science. The US has a long and sad anti-science and anti-intellectual tradition, which seems to be in full flower today.

Exposing science frauds shows the strength of the scientific method. Compare that to politics: Donald Trump's abysmal record of 11 percent True or Mostly True statements - as evaluated by the non-partisan Politifact - and science frauds pale in significance. In comparison, Hillary Clinton has a 50 percent True or Mostly True rating.

I'm hopeful that millennials are embracing an evidence-based worldview. Ideology gives us answers that are simple and satisfying but rarely right.

The real world is messier than ideologues can imagine. Like faked data, slogans are easier, offering the illusion of knowledge without the hard work that understanding requires. It is ultimately up to us consumers to enforce quality in science and in politics.

Courteous comments welcome, of course.